Dynamic gazetteers for entity recognition and fact association

• US 8,429,099 B1
• Filed: 10/14/2010
• Issued: 04/23/2013
• Est. Priority Date: 10/14/2010
• Status: Expired due to Fees

• Current Assignee: ARO, Inc. (Dibona Holding)
• Original Assignee: ARO, Inc. (Dibona Holding)
• Inventors: Perkowitz, Michael, Eustice, Kevin Francis, Lazarus, Jonathan D., Krossa, Kenneth Dwight, Reed, Steven M.
• Primary Examiner(s): CHEN, ALAN S
• Application Number:
  US12/904,957
• Time in Patent Office: 922 Days
• Field of Search: None
• US Class Current: 706/12
• CPC Class Codes: G06F 40/295 Named entity recognition
• CPC Class Codes: G06Q 10/10 Office automation; Time management

Abstract

An entity recognition model is trained to use gazetteers to recognize entities referenced in documents. When a document associated with a user is received for recognizing entities referenced therein, a gazetteer is selected based on user contexts associated with the user. The document is analyzed using the gazetteer according to the entity recognition model to recognize an entity referenced therein. Additional facts associated with the recognized entities and actions applicable to the recognized entities are provided to enable the user to perform the applicable actions against the recognized entity.

Claims

1. A computer-implemented method of recognizing entities referenced in a document, the method comprising:

receiving a document associated with a user;

selecting at least one gazetteer from a plurality of gazetteers based on a context related to the user or the document;

analyzing the document using the selected gazetteer according to an entity recognition model to recognize an entity referenced therein, the entity recognition model having been trained to recognize entities referenced in a training corpus using the plurality of gazetteers as features, wherein the selected gazetteer has been modified responsive to the training; and

retrieving one or more facts associated with the recognized entity for providing the user with applicable actions against the recognized entity using the facts.

2. The method of claim 1, wherein the selected gazetteer includes a plurality of terms associated with a referencing fact used to reference entities of an entity type.

3. The method of claim 1, wherein the selected gazetteer comprises an entity schema for an associated entity type, the entity schema specifying one or more facts associated with entities of the entity type.

4. The method of claim 3, wherein the entity schema comprises facts of one or more entities of the entity type.

5. The method of claim 1, wherein the selected gazetteer includes a filter rule for reducing false positive entity recognition, and wherein the analyzing step comprises:

applying the filter rule for analyzing the document using the selected gazetteer.

6. The method of claim 1, wherein the selected gazetteer has been modified by personalizing the selected gazetteer for the user using personal data of the user after the entity recognition model is trained.

7. The method of claim 1, wherein the selected gazetteer is specified by the user.

8. The method of claim 1, wherein the related context comprises at least one of the following geographic location contexts:

a location of a user device associated with the document and the user, a location appearing in the document, and a location specified by the user, and wherein the selecting step comprises;

selecting a gazetteer associated with one of the geographic location contexts.

9. The method of claim 1, wherein the related context comprises a social context, and wherein the selecting step comprises:

selecting a user-specific gazetteer of the user or a gazetteer associated with an organization affiliated with the user.

10. The method of claim 1, wherein the analyzing step comprises:

parsing the document for a presence of a term in a list of terms in the selected gazetteer; and

responsive for detecting a presence of a term in the list of terms, determining that an entity corresponding to the term is referenced in the document.

11. The method of claim 1, wherein the entity comprises one of the following:

a people type entity, an organization type entity, and a location type entity.

12. The method of claim 1, wherein the selected gazetteer is published by a third party and selected by the user for recognizing entity references in documents associated with the user.

13. The method of claim 12, wherein the selected gazetteer comprises a collection of data comprising names and aliases of a plurality of entities, a plurality of definitions for actions applicable to the plurality of entities, a plurality of filter rules to be applied to possible matches for the names and aliases of the plurality of entities, and a plurality of graphic elements for presenting the plurality of entities.

14. The method of claim 12, wherein the third party avails the selected gazetteer only to users with security credential.

15. A non-transitory computer-readable storage medium storing executable computer program instructions for recognizing entities referenced in a document, the computer program instructions comprising instructions for:

receiving a document associated with a user;

selecting at least one gazetteer from a plurality of gazetteers based on a context related to the user or the document;

analyzing the document using the selected gazetteer according to an entity recognition model to recognize an entity referenced therein, the entity recognition model having been trained to recognize entities referenced in a training corpus using the plurality of gazetteers as features, wherein the selected gazetteer has been modified responsive to the training; and

retrieving one or more facts associated with the recognized entity for providing the user with applicable actions against the recognized entity using the facts.

16. The non-transitory computer-readable storage medium of claim 15, wherein the selected gazetteer includes a plurality of terms associated with a referencing fact used to reference entities of an entity type.

17. The non-transitory computer-readable storage medium of claim 15, wherein the selected gazetteer comprises an entity schema for an associated entity type, the entity schema specifying one or more facts associated with entities of the entity type.

18. The non-transitory computer-readable storage medium of claim 17, wherein the entity schema comprises facts of one or more entities of the entity type.

19. A system for recognizing entities referenced in a document, the system comprising:

a non-transitory computer-readable storage medium comprising computer program code, the computer program code including instructions for;

a device interaction module for receiving a document associated with a user;

a gazetteer selection module for selecting at least one gazetteer from a plurality of gazetteers based on a context related to the user or the document; and

an entity recognition engine configured for analyzing the document using the selected gazetteer to recognize an entity referenced therein, the entity recognition engine having been trained to recognize entities referenced in a training corpus using the plurality of gazetteers as features, the selected gazetteer having been modified responsive to the training, wherein the entity recognition engine is further configured for retrieving one or more facts associated with the recognized entity for providing the user with applicable actions against the recognized entity using the facts; and

one or more processors for executing the computer program code.